Abstract: Objectives: Antarctic coastal polynyas are critical regions for sea ice production and dense water formation. The Terra Nova Bay Polynya (TNBP) is one of the most productive, driven primarily by intense katabatic winds. While its overall variability is known, the specific mechanistic roles of individual atmospheric drivers (wind, temperature, heat fluxes) in controlling daily ice production remain inadequately quantified. This study aims to characterize the spatiotemporal evolution of the TNBP and to elucidate its response mechanisms to atmospheric forcing, specifically focusing on the coupled dynamic and thermodynamic processes over the past two decades. Methods: We utilized MODIS sea ice surface temperature products and ERA5 atmospheric reanalysis data spanning from 2003 to 2023. To address the challenge of frequent cloud cover in polar regions, a three-step fusion cloud removal algorithm was specifically employed to optimize data continuity, significantly reducing the average cloud cover from 45.07% (MOD) and 36.75% (MYD) to 2.85%. A surface energy flux balance model was applied to retrieve thin ice thickness, with the TNBP boundary strictly identified by a thin ice threshold of 0.2 m. Subsequently, we constructed a high-resolution (1 km) dataset of TNBP area and ice production, analyzing spatiotemporal variations through Pearson correlation analysis and detailed case studies of extreme events. Results: A strong positive correlation (R=0.94, p<0.001) was observed between TNBP area and ice production, both exhibiting significant interannual fluctuations. Over the 21-year study period, ice production displayed a significant increasing trend at a rate of 0.99 km³/year (p=0.0076), primarily driven by the expansion of the polynya area rather than an increase in unit area ice production efficiency. This expansion is attributed to the weakening of sea ice physical properties under warming climate, making thinner new ice more susceptible to katabatic winds. Seasonally, both area and ice production followed a unimodal pattern, peaking in March and reaching troughs in August, with the most significant interannual variability observed during autumn (March) and early spring (October). Spatially, areas of high-frequency occurrence and high ice production were consistently concentrated in the coastal waters north of the Drygalski Ice Tongue (DIT), highlighting the synergistic constraints of topographic blocking and katabatic winds. Correlation analysis revealed that latent heat flux exhibited the strongest positive correlation with both TNBP area (0.7775) and ice production (0.8259), indicating its primary role in explaining TNBP's spatiotemporal variability. Katabatic wind speed was the second most influential factor, with correlations of 0.5687 for area and 0.6235 for ice production, underscoring its core dynamic role. Sensible heat flux also showed significant positive correlations (0.5030 for area, 0.5896 for ice production), forming the thermodynamic basis for ice production alongside latent heat flux. Air temperature showed the weakest direct correlation (0.2789 for area, 0.1905 for ice production), suggesting its influence is primarily indirect, modulating sea-air temperature and vapor pressure differences, and highly dependent on wind speed. Case analyses of extreme events confirmed that katabatic wind speed acts as the primary dynamic driver for TNBP development; it determines the extent of open water through mechanical opening and regulates the efficiency of sea-air turbulent exchange. Sensible and latent heat fluxes directly control the ice production rate, with their intensities synergistically modulated by wind speed and air temperature. While air temperature cannot independently drive TNBP changes, it indirectly modifies ice production efficiency by altering sea-air temperature and vapor pressure differences, provided that wind speed exceeds a critical threshold. Conclusions: This study systematically reveals the coupled dynamic-thermodynamic processes governing the TNBP over the past two decades. Katabatic wind speed is identified as the primary dynamic driver, directly determining the extent of open water and indirectly influencing the intensity and stability of sea-air heat exchange by regulating turbulent exchange efficiency. Sensible and latent heat fluxes are direct control factors for ice production, with their release intensity directly dictating sea ice formation efficiency, a process reliant on open interfaces and turbulent conditions provided by the wind field, and modulated by air temperature's effect on sea-air temperature and humidity differences. Air temperature cannot independently drive TNBP changes; its influence on polynya development is indirect, mediated through heat fluxes, and the direction and intensity of this regulatory effect are constrained by wind speed conditions. The high-resolution (1 km) TNBP area and ice production dataset (2003-2023) constructed in this study provides crucial baseline data for understanding AABW formation and sea-ice-atmosphere coupling processes in the Antarctic Ross Sea, and offers observational evidence for simulating Antarctic coastal polynya responses under global climate change.